Skip to content

Capstone: ship a monitored service

Every other topic in this repo is a strong reference — notes plus one isolated, CI-validated example. This one is different: it's a single walkthrough that chains four of those topics together, to show that the pieces actually click — not just that each one works in isolation.

What this is (and isn't)

flowchart LR
    Docker["Docker\nbuild the image"] --> K8s["Kubernetes\ndeploy it"]
    K8s --> Obs["Observability\nscrape it"]
    Docker --> CICD["CI/CD\nautomate the whole chain"]
    K8s --> CICD
    Obs --> CICD
  • Is: one doc that builds topics/docker/examples/hello-nginx, shows the manifests to deploy it to a local kind cluster with an nginx-prometheus-exporter sidecar for a real scraped metric, and the GitHub Actions workflow that would automate all of that on every merge.
  • Isn't: a new category of CI-validated example. It intentionally reuses and adapts topics/docker/, topics/kubernetes/, topics/observability/, and topics/cicd/'s existing examples rather than shipping a fifth copy of the same Dockerfile/Deployment pattern — every manifest here is a small, explained variant of something those topics already cover in depth. There's no examples/ directory under this topic for that reason; go read the topic a step belongs to for the full explanation of why it works, this doc only covers how it connects.

Contents

  • notes/01-ship-a-monitored-service.md — the walkthrough itself.

The Docker build and every manifest's schema have been verified for real; the note's own "what was actually verified" callout says exactly what was and wasn't run live, and why (this piece isn't wired into validate.yml the same way a topic's own examples are — standing up a kind cluster plus a Prometheus scrape target for one doc isn't worth a permanent CI job the way the per-topic live checks are).